带有GARCH误差和趋势项的近单位根过程的估计

近单位根过程在时间序列理论中占有重要地位.因其处于平稳过程与非平稳过程（单位根过程）的中间地带，研究其统计性质的重要价值在于揭示这个中间地带的特殊属性，及其与平稳过程和单位根过程统计属性的差异.在更新项只有二阶矩存在的条件下，讨论了带有GARCH(p,q)误差项与趋势项的近单位根过程的最小二乘估计.并推导了基于最小二乘估计的Dickey-Fuller检验统计量的渐近分布.该结果在单位根检验中具有一定意义.

Estimation of the nearly unit root processes with GARCH errors and a drift

Nearly unit root processes are very important in time series theory. Due to being in the middle field be- tween stationary processes and nonstationary processes, the significant value of nearly unit root processes is that they may uncover the special properties of this field and the difference between the middle field and the other two fields. Least square estimation for nearly unit root processes with GARCH errors and a drift are discussed only un- der finite variance of innovations. And the asymptotic distributions of Dickey-Fuller tests are derived. This result plays certain significance in unit root test.